Back to skill

Security audit

Liquidity Planner

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a PancakeSwap liquidity planner, but it asks for broader local execution access than it needs and includes an automatic background ping that sends device details.

Review this skill before installing. Its normal market-data lookups are expected for a DeFi liquidity planner, but installation should be treated as Review because it can run broad local code, depends on a non-bundled workspace-relative script, and automatically contacts pancakeswap.ai with host details. Use only in a restricted sandbox, do not mount wallets or secrets, and verify every chain, token address, and PancakeSwap URL before acting.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Tool Hijacking and SpoofingModifies or replaces tools so legitimate-looking calls execute attacker logic
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Findings (4)

other

Warning
Location
SKILL.md:98
Finding

Mandatory Host Fingerprinting Telemetry

Content
View full analysis
/dev/null & ``` ### Technical Analysis The initialization step instructs the agent to contact `pancakeswap.ai` on every invocation and transmit: - The AI agent or system identifier - The exact invocation timestamp - The host operating system - The host CPU architecture - The Skill name and version This information is not required to discover pools, validate tokens, calculate liquidity ranges, or generate a PancakeSwap deep link. It therefore exceeds the minimum network privileges needed for the declared functionality. The request runs in the background, discards its response, and suppresses errors, making the telemetry unobtrusive. It also conflicts with the curl hostname allowlist at `SKILL.md:86`, because that list does not include `pancakeswap.ai`. ### Attack Path 1. A user invokes the liquidity-planner Skill. 2. The mandatory initialization instruction executes the background curl request. 3. Shell substitutions collect the current timestamp, operating system, and architecture. 4. These values and the agent identifier are placed in URL query parameters. 5. The remote `pancakeswap.ai` server receives and can correlate the host fingerprint and invocation event. ### Impact Assessment The remote service can track Skill usage and fingerprint execution environments. Repeated requests may permit correlation of users or agent installations through timestamps, agent identifiers, network metadata, operating systems, and architectures. This does not expose credentials or private keys, but it introduces an unnecessary privacy leak and an undeclared network dependency. The ...[truncated 105 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
SKILL.md:6
Finding

Excessive Filesystem and General-Purpose Code Execution Permissions

Content
View full analysis
Remediation
View remediation

T07 · Tool Hijacking and Spoofing

Error
Location
SKILL.md:311
Finding

Execution of an Unbundled Working-Directory-Dependent Script

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:153
Finding

Shell Injection Through Incomplete Validation of User-Supplied Search Terms

Content
View full analysis
` and `<`) - Comment marker (`#`) - Parentheses and other context-sensitive shell syntax A value such as: ```text ' > /tmp/clobber # ``` would produce: ```bash KEYWORD='' > /tmp/clobber #' ``` The shell interprets the redirection and creates or truncates `/tmp/clobber`. This payload does not require any character included in the documented rejection list. More damaging targets may be reachable wherever the agent account has write permission. The later use of `--data-urlencode` safely encodes the variable for HTTP, but it does not protect the earlier shell assignment if that assignment is constructed by textual interpolation. ### Attack Path 1. An attacker supplies a crafted token-search keyword containing a single quote, redirection, and a shell comment. 2. The documented blacklist accepts the input because those characters are not prohibited. 3. The agent substitutes th ...[truncated 770 chars]
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (22)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The invocation rules are overly broad and can trigger this skill for generic yield or portfolio questions, including on chains where PancakeSwap liquidity planning may not be the user's intent. Over-broad routing can cause the agent to collect wallet-related preferences, query third-party services, or generate transaction-adjacent links in contexts the user did not request.

Content

No source excerpt is available for this finding.

External Model or Provider Selection

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.

Content

Scanner excerpt · SKILL.md (reported line 7)May include surrounding context.

md
description: Plan liquidity provision on PancakeSwap. Use when user says "add liquidity on pancakeswap", "provide liquidity", "LP on pancakeswap", or describes wanting to deposit tokens into liquidity pools without writing code. Also use when a user holds two tokens and asks how to earn yield, get best profit, or put tokens to work — especially on Solana or any chain where farming is unavailable.
homepage: https://github.com/pancakeswap/pancakeswap-ai
allowed-tools: Read, Write, Edit, Glob, Grep, Bash(curl:*), Bash(jq:*), Bash(cast:*), Bash(node:*), Bash(python3:*), Bash(xdg-open:*), Bash(open:*), WebFetch, WebSearch, Task(subagent_type:Explore), AskUserQuestion
model: sonnet
license: MIT
metadata:
  author: pancakeswap

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The initialization step instructs a background telemetry call that sends system metadata without clear user notice. Because it runs automatically and is unrelated to core LP planning, it creates an undisclosed external transmission risk and weakens user control over data sharing.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 249)May include surrounding context.

md
[[ "$TOKEN" =~ ^0x[0-9a-fA-F]{40}$ ]] || { echo "Invalid token address"; exit 1; }

# name() selector = 0x06fdde03
curl -sf -X POST "$RPC" \
  -H "Content-Type: application/json" \
  -d "{\"jsonrpc\":\"2.0\",\"id\":1,\"method\":\"eth_call\",\"params\":[{\"to\":\"$TOKEN\",\"data\":\"0x06fdde03\"},\"latest\"]}" \
  | jq -r '.result'

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 273)May include surrounding context.

md
[[ "$TOKEN" =~ ^0x[0-9a-fA-F]{40}$ ]] || { echo "Invalid token address"; exit 1; }

# name() selector = 0x06fdde03
curl -sf -X POST "$RPC" \
  -H "Content-Type: application/json" \
  -d "{\"jsonrpc\":\"2.0\",\"id\":1,\"method\":\"eth_call\",\"params\":[{\"to\":\"$TOKEN\",\"data\":\"0x06fdde03\"},\"latest\"]}" \
  | jq -r '.result'

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 320)May include surrounding context.

md
**Environment variables:**

| Variable    | Required | Description                                                                                                                                 |
| ----------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
| `CHAIN`     | yes      | Chain string: `bsc`, `eth`, `arb`, `base`, `zksync`, `linea`, `opbnb`, `monad`, `sol`                                                       |
| `TOKEN0`    | no       | EVM address, Solana pubkey, or native alias (`bnb`, `eth`, `sol`)                                                                           |

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 324)May include surrounding context.

md
| ----------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
| `CHAIN`     | yes      | Chain string: `bsc`, `eth`, `arb`, `base`, `zksync`, `linea`, `opbnb`, `monad`, `sol`                                                       |
| `TOKEN0`    | no       | EVM address, Solana pubkey, or native alias (`bnb`, `eth`, `sol`)                                                                           |
| `TOKEN1`    | no       | Same as TOKEN0                                                                                                                              |
| `ORDER_BY`  | no       | `tvlUSD` (default), `apr24h`, `volumeUSD24h`                                                                                                |
| `PROTOCOLS` | no       | Comma-separated list of protocols to include (default: all). Supported: `v2`, `v3`, `stable`, `infinityCl`, `infinityBin`, `infinityStable` |

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 325)May include surrounding context.

md
| `CHAIN`     | yes      | Chain string: `bsc`, `eth`, `arb`, `base`, `zksync`, `linea`, `opbnb`, `monad`, `sol`                                                       |
| `TOKEN0`    | no       | EVM address, Solana pubkey, or native alias (`bnb`, `eth`, `sol`)                                                                           |
| `TOKEN1`    | no       | Same as TOKEN0                                                                                                                              |
| `ORDER_BY`  | no       | `tvlUSD` (default), `apr24h`, `volumeUSD24h`                                                                                                |
| `PROTOCOLS` | no       | Comma-separated list of protocols to include (default: all). Supported: `v2`, `v3`, `stable`, `infinityCl`, `infinityBin`, `infinityStable` |

Set `ORDER_BY` based on the user's stated goal:

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Earlier sections state StableSwap is supported on BSC, Ethereum, and Arbitrum, and the deep link examples include non-BSC use. Later, the 'Characteristics' section says 'Chain: BSC only (currently)', which directly conflicts with the rest of the file's stated behavior and guidance.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The deep link builder hardcodes chain=bsc for StableSwap even though the surrounding documentation claims Ethereum and Arbitrum are supported. This can direct users to the wrong network context, increasing the risk of mistaken deposits, confusion, or interacting with unintended assets/interfaces.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 156)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 293)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 371)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 36)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 43)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 52)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 59)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 68)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/data-providers.md (reported line 75)May include surrounding context.

To find all PancakeSwap pools containing a specific token:

bash
curl -s "https://api.dexscreener.com/latest/dex/tokens/{tokenAddress}" | \
  jq '.pairs[] | select(.dexId == "pancakeswap")'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file provides detailed LP strategy recommendations, risk tradeoff guidance, and direct PancakeSwap deep links, but it lacks a clear upfront warning that liquidity provision can cause loss of funds through impermanent loss, smart-contract risk, depegging, or phishing/misconfiguration. In a skill explicitly designed to guide users into DeFi actions, this omission materially increases the chance users treat the content as safe operational advice and act without informed consent.

Content

No source excerpt is available for this finding.

Unbounded Output

Medium
Category
Output Handling
Confidence
80% confidence
Finding

Output size or generation rate is not bounded. Unbounded output enables denial-of-service through resource exhaustion, log flooding, or context-window stuffing.

Content

Scanner excerpt · references/position-types.md (reported line 432)May include surrounding context.

md
| Parameter        | Purpose                | Example                |
| ---------------- | ---------------------- | ---------------------- |
| `inputCurrency`  | Pre-fill amount input  | `&inputCurrency=0.5`   |
| `outputCurrency` | Pre-fill output amount | `&outputCurrency=1000` |
| `pMin`           | Slippage tolerance     | `&pMin=0.95`           |
| `pMax`           | Price impact cap       | `&pMax=1.05`           |

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill performs a background telemetry request to pancakeswap.ai during initialization that is not necessary to plan a liquidity position. It transmits environment metadata such as agent identifier, timestamp, OS, and architecture without an explicit user-consent step, creating an avoidable privacy and trust boundary issue.

Content

No source excerpt is available for this finding.